jian he wrote: > CREATE > DATABASE dbicu1 LOCALE_PROVIDER icu LOCALE 'en_US.UTF-8' ICU_LOCALE > 'en-u-kf-upper' TEMPLATE 'template0'; > CREATE DATABASE dbicu2 LOCALE_PROVIDER icu LOCALE 'en_US.UTF-8' ICU_LOCALE > 'en-u-kr-latn-digit' TEMPLATE 'template0'; > [...] > I am not sure this is my personal misunderstanding. > In the above examples, the first character of column *test_kr* > is so different that the comparison is based on the first letter. > If the first letter is the same then compute the second letter.. > So for whatever collation, I should expect 'A 19' to be adjacent with 'A > 11'?
The query "SELECT test_kr FROM icu ORDER BY def;" does not order by test_kr, so the contents of test_kr have no bearing on the order of the results. If you order by test_kr, the results look like what you're expecting: dbicu1=# SELECT test_kr,def FROM icu ORDER BY test_kr; test_kr | def ---------+----- 1 a | a 8 p | B A 11 | b A 19 | A a 7 | a Œ 1 | a dbicu2=# SELECT test_kr,def FROM icu ORDER BY test_kr ; test_kr | def ---------+----- A 11 | b A 19 | A a 7 | a Œ 1 | a 1 a | a 8 p | B Best regards, -- Daniel Vérité https://postgresql.verite.pro/ Twitter: @DanielVerite
